Understanding systematic reviews and meta-analysis

Free

Loading

  1. A K Akobeng
  1. Correspondence to:
    Dr A K Akobeng
    Department of Paediatric Gastroenterology, Central Manchester and Manchester Children's University Hospitals, Berth Hall Children'south Hospital, Charlestown Road, Blackley, Manchester, M9 7AA, Great britain; tony.akobengcmmc.nhs.uk

Abstruse

This review covers the basic principles of systematic reviews and meta-analyses. The bug associated with traditional narrative reviews are discussed, as is the function of systematic reviews in limiting bias associated with the assembly, disquisitional appraisal, and synthesis of studies addressing specific clinical questions. Of import issues that demand to be considered when appraising a systematic review or meta-analysis are outlined, and some of the terms used in the reporting of systematic reviews and meta-analyses—such as odds ratio, relative risk, confidence interval, and the woods plot—are introduced.

  • RCT, randomised controlled trial
  • systematic review
  • meta-analysis
  • narrative review
  • critical appraisal

Statistics from Altmetric.com

  • RCT, randomised controlled trial
  • systematic review
  • meta-analysis
  • narrative review
  • critical appraisement

Health intendance professionals are increasingly required to base their practice on the best available evidence. In the outset article of the series, I described bones strategies that could be used to search the medical literature.1 After a literature search on a specific clinical question, many articles may be retrieved. The quality of the studies may be variable, and the individual studies might have produced conflicting results. Information technology is therefore important that health intendance decisions are not based solely on i or ii studies without account being taken of the whole range of research information available on that topic.

Health care professionals take always used review manufactures as a source of summarised evidence on a item topic. Review articles in the medical literature take traditionally been in the grade of "narrative reviews" where experts in a particular field provide what is supposed to be a "summary of testify" in that field. Narrative reviews, although all the same very mutual in the medical field, take been criticised because of the loftier risk of bias, and "systematic reviews" are preferred.ii Systematic reviews utilise scientific strategies in ways that limit bias to the assembly, a critical appraisement, and synthesis of relevant studies that address a specific clinical question.ii

THE Problem WITH TRADITIONAL REVIEWS

The validity of a review article depends on its methodological quality. While traditional review articles or narrative reviews tin can be useful when conducted properly, there is testify that they are usually of poor quality. Authors of narrative reviews ofttimes use informal, subjective methods to collect and translate studies and tend to be selective in citing reports that reinforce their preconceived ideas or promote their own views on a topic.3, four They are also rarely explicit well-nigh how they selected, assessed, and analysed the primary studies, thereby not assuasive readers to assess potential bias in the review process. Narrative reviews are therefore often biased, and the recommendations fabricated may be inappropriate.five

WHAT IS A SYSTEMATIC REVIEW?

In dissimilarity to a narrative review, a systematic review is a course of research that provides a summary of medical reports on a specific clinical question, using explicit methods to search, critically assess, and synthesise the world literature systematically.6 Information technology is peculiarly useful in bringing together a number of separately conducted studies, sometimes with conflicting findings, and synthesising their results.

By providing in a clear explicit style a summary of all the studies addressing a specific clinical question,iv systematic reviews let us to take business relationship of the whole range of relevant findings from research on a particular topic, and not just the results of i or two studies. Other advantages of systematic reviews have been discussed by Mulrow.7 They tin be used to establish whether scientific findings are consequent and generalisable across populations, settings, and treatment variations, or whether findings vary significantly past particular subgroups. Moreover, the explicit methods used in systematic reviews limit bias and, hopefully, will meliorate reliability and accuracy of conclusions. For these reasons, systematic reviews of randomised controlled trials (RCTs) are considered to be prove of the highest level in the hierarchy of research designs evaluating effectiveness of interventions.8

METHODOLOGY OF A SYSTEMATIC REVIEW

The demand for rigour in the preparation of a systematic review means that in that location should exist a formal process for its deport. Figure i summarises the process for conducting a systematic review of RCTs.ix This includes a comprehensive, exhaustive search for primary studies on a focused clinical question, pick of studies using clear and reproducible eligibility criteria, critical appraisal of primary studies for quality, and synthesis of results according to a predetermined and explicit method.3, ix

WHAT IS A META-Analysis?

Post-obit a systematic review, information from private studies may be pooled quantitatively and reanalysed using established statistical methods.ten This technique is called meta-assay. The rationale for a meta-analysis is that, by combining the samples of the individual studies, the overall sample size is increased, thereby improving the statistical ability of the assay likewise every bit the precision of the estimates of treatment furnishings.11

Meta-analysis is a two stage procedure.12 The offset stage involves the calculation of a measure of treatment effect with its 95% confidence intervals (CI) for each private report. The summary statistics that are usually used to measure treatment upshot include odds ratios (OR), relative risks (RR), and run a risk differences.

In the second stage of meta-assay, an overall treatment effect is calculated as a weighted boilerplate of the individual summary statistics. Readers should note that, in meta-analysis, data from the individual studies are not simply combined as if they were from a single written report. Greater weights are given to the results from studies that provide more data, because they are likely to be closer to the "truthful effect" we are trying to estimate. The weights are frequently the changed of the variance (the square of the standard fault) of the handling effect, which relates closely to sample size.12 The typical graph for displaying the results of a meta-analysis is called a "forest plot".xiii

The forest plot

The plot shows, at a glance, information from the individual studies that went into the meta-analysis, and an approximate of the overall results. It as well allows a visual assessment of the corporeality of variation between the results of the studies (heterogeneity). Figure 2 shows a typical wood plot. This figure is adapted from a contempo systematic review and meta-analysis which examined the efficacy of probiotics compared with placebo in the prevention and treatment of diarrhoea associated with the use of antibiotics.fourteen

Description of the forest plot

In the woods plot shown in fig 2, the results of nine studies have been pooled. The names on the left of the plot are the first authors of the chief studies included. The black squares correspond the odds ratios of the private studies, and the horizontal lines their 95% confidence intervals. The area of the black squares reflects the weight each trial contributes in the meta-assay. The 95% conviction intervals would contain the truthful underlying effect in 95% of the occasions if the study was repeated again and again. The solid vertical line corresponds to no effect of handling (OR = 1.0). If the CI includes 1, then the difference in the issue of experimental and control treatment is not meaning at conventional levels (p>0.05).15 The overall handling effect (calculated as a weighted average of the individual ORs) from the meta-analysis and its CI is at the lesser and represented as a diamond. The centre of the diamond represents the combined handling outcome (0.37), and the horizontal tips represent the 95% CI (0.26 to 0.52). If the diamond shape is on the Left of the line of no effect, and then Less (fewer episodes) of the result of interest is seen in the treatment group. If the diamond shape is on the Right of the line, then moRe episodes of the outcome of interest are seen in the treatment grouping. In fig 2, the diamond shape is institute on the left of the line of no effect, meaning that less diarrhoea (fewer episodes) was seen in the probiotic grouping than in the placebo grouping. If the diamond touches the line of no issue (where the OR is 1) then there is no statistically pregnant departure betwixt the groups being compared. In fig ii, the diamond shape does not touch the line of no effect (that is, the confidence interval for the odds ratio does non include i) and this ways that the deviation found between the ii groups was statistically meaning.

APPRAISING A SYSTEMATIC REVIEW WITH OR WITHOUT META-ANALYSIS

Although systematic reviews occupy the highest position in the hierarchy of testify for manufactures on effectiveness of interventions,8 it should non be causeless that a study is valid simply because it is stated to be an systematic review. Merely as in RCTs, the main bug to consider when appraising a systematic review can be condensed into three important areas8:

  • The validity of the trial methodology.

  • The magnitude and precision of the handling event.

  • The applicability of the results to your patient or population.

Box 1 shows a listing of ten questions that may exist used to appraise a systematic review in all three areas.16

Box ane: Questions to consider when appraising a systematic review16

  • Did the review address a clearly focused question?

  • Did the review include the correct type of written report?

  • Did the reviewers try to identify all relevant studies?

  • Did the reviewers assess the quality of all the studies included?

  • If the results of the written report have been combined, was it reasonable to do then?

  • How are the results presented and what are the chief results?

  • How precise are the results?

  • Can the results be applied to your local population?

  • Were all important outcomes considered?

  • Should practice or policy change as a result of the evidence independent in this review?

ASSESSING THE VALIDITY OF TRIAL METHODOLOGY

Focused enquiry question

Similar all research reports, the authors should clearly state the research question at the outset. The research question should include the relevant population or patient groups being studied, the intervention of interest, any comparators (where relevant), and the outcomes of involvement. Keywords from the research question and their synonyms are usually used to place studies for inclusion in the review.

Types of studies included in the review

The validity of a systematic review or meta-analysis depends heavily on the validity of the studies included. The authors should explicitly country the type of studies they have included in their review, and readers of such reports should decide whether the included studies have the appropriate study blueprint to answer the clinical question. In a recent systematic review which determined the effects of glutamine supplementation on morbidity and weight gain in preterm babies the investigators based their review just on RCTs.17

Search strategy used to identify relevant articles

There is bear witness that single electronic database searches lack sensitivity and relevant manufactures may be missed if only one database is searched. Dickersin et al showed that only 30–80% of all known published RCTs were identifiable using MEDLINE.xviii Even if relevant records are in a database, it tin can be difficult to call up them easily. A comprehensive search is therefore important, not only for ensuring that every bit many studies equally possible are identified but also to minimise selection bias for those that are found. Relying exclusively on one database may call up a prepare of studies that are unrepresentative of all studies that would take been identified through a comprehensive search of multiple sources. Therefore, in club to retrieve all relevant studies on a topic, several dissimilar sources should be searched to identify relevant studies (published and unpublished), and the search strategy should not be limited to the English linguistic communication. The aim of an extensive search is to avoid the problem of publication bias which occurs when trials with statistically pregnant results are more than likely to be published and cited, and are preferentially published in English language journals and those indexed in Medline.

In the systematic review referred to in a higher place, which examined the furnishings of glutamine supplementation on morbidity and weight gain in preterm babies, the authors searched the Cochrane controlled trials register, Medline, and Embase,17 and they also manus searched selected journals, cross referencing where necessary from other publications.

Quality assessment of included trials

The reviewers should state a predetermined method for assessing the eligibility and quality of the studies included. At to the lowest degree two reviewers should independently assess the quality of the included studies to minimise the take chances of selection bias. There is evidence that using at to the lowest degree ii reviewers has an important effect on reducing the possibility that relevant reports volition be discarded.19

Pooling results and heterogeneity

If the results of the individual studies were pooled in a meta-assay, it is of import to determine whether information technology was reasonable to practice and then. A clinical judgement should exist made nearly whether it was reasonable for the studies to be combined based on whether the individual trials differed considerably in populations studied, interventions and comparisons used, or outcomes measured.

The statistical validity of combining the results of the various trials should be assessed by looking for homogeneity of the outcomes from the various trials. In other words, in that location should be some consistency in the results of the included trials. One style of doing this is to inspect the graphical display of results of the private studies (forest plot, meet above) looking for similarities in the management of the results. When the results differ greatly in their direction—that is, if there is meaning heterogeneity—then it may not be wise for the results to be pooled. Some manufactures may also report a statistical test for heterogeneity, but information technology should be noted that the statistical power of many meta-analyses is usually too low to let the detection of heterogeneity based on statistical tests. If a study finds pregnant heterogeneity among reports, the authors should attempt to offer explanations for potential sources of the heterogeneity.

Magnitude of the treatment outcome

Mutual measures used to written report the results of meta-analyses include the odds ratio, relative run a risk, and mean differences. If the outcome is binary (for instance, disease v no affliction, remission v no remission), odds ratios or relative risks are used. If the outcome is continuous (for example, claret pressure level measurement), mean differences may be used.

ODDS RATIOS AND RELATIVE RISKS

Odds and odds ratio

The odds for a group is defined equally the number of patients in the group who achieve the stated finish point divided by the number of patients who do not. For example, the odds of acne resolution during treatment with an antibiotic in a group of 10 patients may be 6 to four (6 with resolution of acne divided by 4 without  =  1.5); in a control grouping the odds may be 3 to vii (0.43). The odds ratio, as the proper name implies, is a ratio of 2 odds. Information technology is just defined every bit the ratio of the odds of the treatment grouping to the odds of the control group. In our example, the odds ratio of treatment to command group would exist 3.5 (1.5 divided past 0.43).

Risk and relative risk

Risk, as opposed to odds, is calculated every bit the number of patients in the grouping who achieve the stated terminate bespeak divided past the total number of patients in the group. Risk ratio or relative risk is a ratio of ii "risks". In the example above the risks would be 6 in 10 in the treatment group (6 divided by 10  =  0.6) and three in x in the command group (0.three), giving a run a risk ratio, or relative risk of ii (0.6 divided past 0.3).

Interpretation of odds ratios and relative risk

An odds ratio or relative gamble greater than 1 indicates increased likelihood of the stated outcome being achieved in the treatment group. If the odds ratio or relative risk is less than 1, there is a decreased likelihood in the treatment group. A ratio of 1 indicates no difference—that is, the effect is merely every bit likely to occur in the treatment grouping as information technology is in the control group.11 As in all estimates of treatment effect, odds ratios or relative risks reported in meta-analysis should be accompanied by conviction intervals.

Readers should sympathize that the odds ratio volition be close to the relative risk if the end point occurs relatively infrequently, say in less than 20%.15 If the consequence is more than common, and then the odds ratio volition considerably overestimate the relative risk. The advantages and disadvantages of odds ratios v relative risks in the reporting of the results of meta-analysis take been reviewed elsewhere.12

Precision of the treatment consequence: conviction intervals

As stated earlier, confidence intervals should accompany estimates of handling effects. I discussed the concept of conviction intervals in the second article of the series.8 Ninety v per cent confidence intervals are commonly reported, only other intervals such as 90% or 99% are also sometimes used. The 95% CI of an guess (for example, of odds ratios or relative risks) will exist the range within which we are 95% certain that the true population treatment effect will lie. The width of a conviction interval indicates the precision of the approximate. The wider the interval, the less the precision. A very long interval makes us less sure nearly the accuracy of a study in predicting the truthful size of the effect. If the confidence interval for relative risk or odds ratio for an estimate includes 1, then nosotros take been unable to demonstrate a statistically significant difference between the groups being compared; if it does not include 1, then we say that at that place is a statistically significant departure.

APPLICABILITY OF RESULTS TO PATIENTS

Health intendance professionals should always make judgements about whether the results of a item study are applicable to their own patient or group of patients. Some of the issues that one demand to consider before deciding whether to contain a particular piece of research show into clinical do were discussed in the second commodity of the series.viii These include similarity of study population to your population, benefit v harm, patients preferences, availability, and costs.

CONCLUSIONS

Systematic reviews apply scientific strategies to provide in an explicit fashion a summary of all studies addressing a specific question, thereby allowing an account to be taken of the whole range of relevant findings on a item topic. Meta-analysis, which may accompany a systematic review, can increase power and precision of estimates of treatment effects. People working in the field of paediatrics and child health should understand the cardinal principles of systematic reviews and meta-analyses, including the ability to apply critical appraisement not only to the methodologies of review articles, but as well to the applicability of the results to their ain patients.

REFERENCES

Asking Permissions

If you wish to reuse any or all of this article please utilize the link below which will take yous to the Copyright Clearance Center's RightsLink service. You lot volition exist able to become a quick cost and instant permission to reuse the content in many dissimilar ways.